Loading dose plus continuous/extended infusion versus intermittent bolus of β-lactams for the treatment of Gram-negative bacteria bloodstream infections: a propensity score-adjusted retrospective cohort study

Abstract

Abstract Background Optimal β-lactam dosing for the treatment of Gram-negative bacteria bloodstream infections (GNB-BSIs) remains a debated issue. Herein, the efficacy and safety of a loading dose (LD) followed by extended/continuous infusion (EI/CI) versus intermittent bolus (IB) of these drugs for the treatment of GNB-BSIs was evaluated. Methods This is a retrospective observational study enrolling patients with GNB-BSIs treated with β-lactams from 1 October 2020 to 31 March 2022. The 30 day infection-related mortality rate was assessed with Cox regression, while mortality risk reduction was evaluated by an inverse probability of treatment weighting regression adjustment (IPTW-RA) model. Results Overall, 224 patients were enrolled: 140 and 84 in the IB and EI/CI groups, respectively. β-Lactam regimens were chosen according to pathogen antibiogram, clinical judgement and current guidelines. Interestingly, the LD + EI/CI regimen was associated with a significant lower mortality rate (17% versus 32%, P = 0.011). Similarly, β-lactam LD + EI/CI was significantly associated with a reduced risk of mortality at multivariable Cox regression [adjusted HR (aHR) = 0.46; 95%CI = 0.22–0.98; P = 0.046]. Finally, the IPTW-RA (adjusted for multiple covariates) was performed, showing a significant risk reduction in the overall population [−14% (95% CI = −23% to −5%)]; at the subgroup restricted analysis, a significant risk reduction (>15%) was observed in the case of GNB-BSI in severely immunocompromised patients (P = 0.003), for SOFA score > 6 (P = 0.014) and in septic shock (P = 0.011). Conclusions The use of LD + EI/CI of β-lactams in patients with a GNB-BSI may be associated with reduced mortality; also in patients with severe presentation of infection or with additional risk factors, such as immunodepression.